Chantix and Side Effects

WSJ's Health Blog devotes a post to some of the craziness surrounding the side effects that are being reported all of a sudden.

In case you've not been paying attention, there have been two waves of adverse event reports starting with reports of neuropsychiatric effects, such as suicide and depression. Now there are reports of convulsions in drivers and even simple personal accidents. The FAA banned pilots from taking the drug and sales of Chantix (varenicline) are collapsing.

I am not a fan of big pharma and only rarely do I become a grudging apologist, but I will take a healthy skeptical approach to the situation and shoot from the comfort of my lounge chair.

Part of this tempest is a study by the Institute For Safe Medicine Practices, which is involved in a new pilot method of reporting adverse events. As far as I understand, the pilot involves quarterly analysis of the Adverse Event Reporting (AER) Database, which accepts electronic side effect reports.

All adverse events are recorded and linked to a drug. At the point that a report, or any number of reports are made, there can be no assumption of causation. There are a set of epidemiological principles for establishing causation out of associations.

There is an additional problem that is well-known with voluntary reporting mechanisms... the same thing that customer service people know. A few people who squawk the most have the potential of skewing your results. Electronic reporting should make reporting more complete and thus less susceptible to the bias introduced by voluntary reporting. On the other hand, one could hypothesize that electronic reporting amplifies distortions in public perception.

Once a few reports are made, such that an association is reported in the media, things may snowball. People may begin report minimal events that they would not have otherwise reported or events that they did not consider to report. The media is incapable of shading nuances of association and causation or the real meaning of a further investigation by the FDA. Once reports exist, the FDA and the ISMP have no choice but to issue warning statements.

That does not mean varenicline is a bad drug, nor that it is more unsafe than continuing to smoke. It does mean that a new monitoring system is functioning as designed (assuming it is not being manipulated or introducing worse biases than it was intended to prevent.) We will have more information about adverse events for new drugs sooner and more accurately, for the benefit of empowered consumers.

Why should sales go down?